A massive analysis of census and tax records indicates that inherited genetic factors account for nearly one-third of the variation in earnings among steadily employed Americans in their late 20s and 30s. At the same time, the research suggests that standard twin studies may overstate the genetic component of career success, partly because same-sex twins are unusually likely to end up working at the same company. The findings were published in PNAS.
Why do children from certain families consistently earn higher paychecks in adulthood than children from other families? Social scientists examine this question using a statistical concept called heritability, which measures the proportion of differences in a trait across a population that can be traced to inherited genetic variation rather than environmental conditions.
To estimate heritability without directly sequencing DNA, researchers have traditionally relied on the classical twin design. This method compares identical twins, who develop from a single fertilized egg and share 100 percent of their genes, with fraternal twins, who develop from two separate eggs and share about 50 percent of their genes, just like ordinary brothers and sisters.
Because both types of twins grow up in the same household, researchers attribute any extra similarity between identical twins to their shared DNA. This rests on what is known as the equal-environments assumption, which holds that identical and fraternal twins share their home environment to the same degree.
Sorting out genetic inheritance from family upbringing has proven challenging in economics. For instance, a 2008 study of American brothers and sisters found that siblings’ adult earnings are moderately correlated, indicating that family background accounts for a sizable share of economic inequality. But ordinary sibling comparisons cannot separate shared DNA from parental wealth and parenting styles.
Twins offer a way to separate those channels. A 2012 review of genetics and economics indicated that long-term income appears roughly as heritable as personality traits or many medical conditions. The review’s authors also warned that separating true genetic effects from social influences remains difficult.
Complicating matters further, a person’s paycheck is not determined solely by the individual skills, education, or personality they bring to the labor market. Labor economists separate wages into two main parts. The first is a portable worker component, which reflects an individual’s earning capacity across different jobs. The second is an employer pay premium, the extra pay or discount a specific company gives to all of its staff.
In a 2013 study of West German workers, study co-author David Card and colleagues found that growing differences between workers and in company pay premiums, along with the increasing sorting of high-earning workers into high-paying companies, accounted for much of the rise in wage inequality. Whether access to these high-paying companies reflects inherited traits, personal connections, or simple luck has remained an open question.
Because population registries rarely record whether twins are identical or fraternal, previous twin studies of income have typically relied on small survey samples. To investigate how genetics and social interactions shape both individual earning power and workplace pay premiums on a national scale, Card and Jesse Rothstein of the University of California, Berkeley, along with Moises Yi of the US Census Bureau, analyzed administrative records covering the full population of young adult twins in the United States.
Unlike physical traits such as height, career outcomes depend heavily on social choices, including where a person wants to work and who else works there. The authors reasoned that if twins influence each other’s job choices or prefer working together at the same workplace, standard twin comparisons could mistake that social coordination for genetic inheritance.
The researchers linked household records from the 2000 United States Census to quarterly wage records from state unemployment insurance tax filings. These records cover formal wage and salary jobs, and the team used data from 2018 through the first quarter of 2023. They omitted the last three quarters of 2020 to avoid labor market disruptions caused by the COVID-19 pandemic.
The sample tracked biological children of household heads who were between 9 and 17 years old in 2000 (born between 1983 and 1990). Twins were defined as children in the same household born within two days of each other, and non-twin siblings as children born at least nine months apart. The employment analysis encompassed 87,000 twin pairs (60,000 same-sex and 27,000 opposite-sex) and 2,168,000 non-twin sibling pairs (1,072,000 same-sex and 1,096,000 opposite-sex).
For the detailed wage analysis, the authors focused on pairs in which both siblings maintained stable full-time work. This meant having at least eight quarters in which they earned more than $3,800 from one employer (roughly full-time minimum-wage pay), excluding the first and last quarters of each job spell. That restriction yielded 44,500 twin pairs (31,500 same-sex and 13,000 opposite-sex) and 1,007,000 non-twin sibling pairs (507,000 same-sex and 500,000 opposite-sex).
Because census files do not record whether twins are identical or fraternal, the researchers relied on biological sex. All opposite-sex twins are fraternal, whereas same-sex twins are a mix of fraternal and identical twins. Identical twins made up an estimated 58.7 percent of same-sex pairs in the sample.
Families may also treat sons and daughters differently. To filter out that possibility, the team used a statistical technique called difference-in-differences. They subtracted the similarity gap between same-sex and opposite-sex regular siblings from the similarity gap between same-sex and opposite-sex twins.
The team analyzed average quarterly earnings using a logarithmic scale, a standard economic measure that captures proportional differences in pay. They also used a statistical wage decomposition model to split each worker’s pay into two parts: their portable individual earning capacity and the average pay premium of the companies they worked for. The model controlled for work experience and calendar time.
That employer pay premium was further broken down into three pieces. These were a geographic location premium (the average employer pay premium among companies in the local commuting zone), an industry premium (the average employer pay premium among companies in that industry), and a remaining workplace-specific pay premium.
Across the full wage sample, same-sex twins showed a correlation in average earnings of 0.543 (on a scale where 0 indicates no relationship and 1 indicates that one twin’s earnings perfectly predict the other’s). By comparison, the correlation was 0.317 for opposite-sex twins, 0.386 for same-sex non-twin siblings, and 0.265 for opposite-sex non-twin siblings.
The researchers also examined non-twin siblings separated by age gaps of one to three years, excluding pairs born less than a year apart to avoid adoption or recording errors. Earnings similarity barely faded as the age difference grew, and the gap between same-sex and opposite-sex siblings remained steady. This pattern supported the assumption that shared family and gender influences do not change drastically for siblings born close together.
Based on their extended twin model, the researchers initially estimated that inherited genetic factors accounted for 36 percent of the variance in overall earnings. Genetic factors also accounted for an estimated 23 percent of the variance in having any employment and 32 percent of the variance in the number of quarters employed.
When the researchers separated earnings into individual and workplace components, portable individual earning power had an estimated heritability of 40 percent. The employer pay premium had an estimated heritability of 21 percent.
Within the workplace component, the industry pay premium had a heritability estimate of 26 percent and the remaining workplace-specific premium had an estimate of 20 percent. The geographic location premium (18 percent) was not statistically significant, because most siblings remained in the region where they grew up.
Finding heritability in employer pay premiums even within regions and industries made it look on the surface as though genetic factors affected not just workers’ skills, but also their access to high-paying companies. Because the 21 percent estimate was roughly twice as large as what skill-based hiring alone would produce, the authors examined whether twins were ending up at the same establishments.
The employment records showed that 11.4 percent of same-sex twin pairs worked at the exact same establishment at some point between 2018 and 2023 (not necessarily at the same time). That was double the rate for same-sex non-twin siblings (5.7 percent), triple the rate for opposite-sex twins (3.8 percent), and more than five times the rate for opposite-sex non-twin siblings (2.2 percent).
The researchers then re-estimated their model after removing all twin and sibling pairs who ever worked at the same establishment. The estimated heritability of the employer pay premium dropped by two-thirds, from 21 percent to 7 percent, and was no longer statistically significant. The heritability of the industry premium fell from 26 percent to 12 percent, and that of the workplace-specific premium fell from 20 percent to essentially zero.
The remaining link between genetics and employer pay premiums fit what would be expected from high-paying companies hiring high-skilled workers alone. Excluding coworkers also lowered the estimated heritability of portable individual earning power from 40 percent to 35 percent. The estimated heritability of overall earnings fell from 36 percent to 31 percent (a relative decline of about 15 percent), and the authors regard these lower figures as their best estimates.
Even with that downward adjustment, the finding that genetics account for roughly a third of individual earning differences aligns with a recent German study, which found that higher cognitive ability at age 23 predicted higher socioeconomic status at age 27 largely due to shared genetic factors. That study looked at educational attainment and occupational prestige in young adulthood rather than wages, but both point to a substantial genetic influence on socioeconomic outcomes.
“The results show that estimated heritability is clearly present, as one would totally expect, but the effect depends a lot on whether the twins work at the same firms,” said Felix Bittmann, a postdoctoral researcher at the Leibniz Institute for Educational Trajectories and the University of Bamberg who was not involved in the research. “It’s an interesting aspect that deserves more attention. Clearly, the paper shows that even well-done genetically informed studies cannot partial out ‘pure’ genetic effects as social interactions will always be present and can influence the findings strongly.”
“While this study is done with twins, its implications extend a lot further: in the social sciences, it is an important question to understand how success is determined (nature vs. nurture),” Bittmann noted. “Twin designs are usually an elegant way to attempt to account for both factors, which classical research designs (surveys) cannot capture at all, as genetic transmission pathways are usually not measured; often, even proxies are lacking (like IQ tests for both children and parents).”
Sociologists often try to separate what they call primary and secondary effects: how family background shapes a child’s academic performance, and how families make choices about continuing education. Those two pathways are harder to separate when genetic inheritance goes unmeasured.
“Especially in sociology, when primary and secondary effects are to be separated, genetic pathways are usually ignored, which is a biasing factor,” Bittmann said. “This research can help to estimate how large such a bias could probably be.”
“The research design is clever, as they exploit a huge US database with official statistics,” Bittmann told PsyPost. “These databases do not contain information about twins, so this information is lacking, but they can reconstruct it from birth dates, which is a good idea. While they need to make some assumptions and correct for the aspect that same-sex twins can be identical or not …, they can account for this statistically; I think this part is fine.”
As with all research, there are a few things to keep in mind. A heritability estimate of 31 percent describes variation across a population. It does not mean that 31 percent of any one person’s paycheck comes from their genes, or that an individual’s earnings are biologically fixed at birth.
Even when genetic differences influence earnings, they can operate through indirect or social pathways, such as how parents’ genetic traits shape the home environment they create for their children. In addition, the wage analysis only included sibling pairs in which both people held stable, full-time wage and salary jobs. That leaves out self-employed workers, federal employees, and adults with minimal attachment to the labor market.
Model assumptions can also pull the heritability estimates in opposite directions. The baseline model assumes parents pair up randomly with respect to genes related to earnings. If spouses instead tend to share genetic traits tied to education and earning potential (known as assortative mating), regular siblings and fraternal twins would share more than half of those relevant genes. The authors calculate this could make the estimates about 16 percent higher than reported (for example, roughly 36 percent rather than 31 percent).
On the other hand, even after removing siblings who worked at the same company, the remaining estimates for portable earning power may still be inflated by unmeasured social interactions. Twins might, for instance, influence each other’s education choices or decide to attend the same college.
The authors suggest this kind of sibling influence could be one contributor to the “missing heritability” problem in genetics. Studies that directly analyze DNA using polygenic scores (scores that add up the small effects of many genetic variants) estimate lower heritability for social outcomes like education and income than twin studies do.
“I’d say the paper confirms known aspects with a larger and better sample size and is generally trustworthy,” Bittmann said. “When it comes to disentangling nature vs. nurture, the paper contributes but can, of course, not give a definite answer. Maybe other researchers can use the same research plan on other databases from other countries to add more insight and external validity.”
The study, “Heritability and social interaction in labor market outcomes: Evidence from the population of twins in the United States,” was authored by David Card, Jesse Rothstein, and Moises Yi.